An Assessment of the Effectiveness of Complex Alternatives in Multiple Choice Achievement Test Items

Abstract

Discrimination indices and difficulty levels of multiple-choice achievement test items containing only substantive response alternatives were compared with items containing each of three complex alternative types: All of the above; None of the above; and combination complex alternatives (e.g., A and B; Either A or B; Two of the above, etc.). The same item statistics were compared, with items reclassified according to type of alternative keyed as the correct answer. Items containing combination complex alternatives were found to be most difficult, and items containing an "All of the above" alternative were found to be least difficult (especially when that alternative was keyed as the correct answer). Discrimination index was less affected by the inclusion or exclusion of complex alternatives than was difficulty level, but the highest discrimination occurred in items containing only substantive alternatives; the lowest in items in which "None of the above" was keyed as the correct answer. It was also found that all three types of complex alternatives functioned better as distractors than did substantive alternatives, with combination complex distractors receiving the highest rate of response.